March 29th, 2022
Businesses looking to auto-write short ads, emails, product descriptions, social media posts and similar copy can now do so in seconds; thanks to new AI writing technology GPT-3.
Developed by OpenAI — a group of leading edge AI researchers that received $1 billion in funding from Microsoft — GPT-3 promises to significantly automate all forms of commercial writing in coming years.
Already, GPT serves as the core engine for dozens upon dozens of AI writing software solutions and other NLP based solutions that license its wizardry — which are available to companies and others for as little as $19/month.
These solutions essentially ‘sit atop’ GPT-3, mostly serving as interfaces that control the underlying GPT-3 auto-writing engine.
While all of these software solutions have only cropped-up in little more than a couple of years, many are already attempting to corner the market by specializing in particular forms of AI-generated writing.
An AI expert, of course, can expertly navigate you through the myriad choices available in this genre of AI writing to ensure you choose the perfect fit for your business or organization.
Marketers looking for help with social media messaging or newsletters, for example, will want to check-out GPT-3 powered auto-writing services that ensure copy is optimized for the gender, age, shopping preferences and other demographics of their target market.
Other marketers who are especially concerned with search engine optimization (SEO) will find a number of GPT-3 based auto-writing services that auto-produce copy optimized for search.
GPT-3 powered auto-writers can accommodate users who are looking to mimic the writing style of a specific newspaper or magazine, render copy that exhibits a certain tone — such as ‘convincing,’ ‘awestruck’ and ‘inspirational — or use data from the Web to personalize messaging that targets specific individuals, companies or organizations.
Moreover, users looking for a Swiss Army Knife writing solution that ‘does it all’ will find GPT-3 based auto-writers designed to handle 50 or more auto-writing tasks.
Available for use by select AI programmers since Summer 2020, GPT-3 differs from earlier versions in that it relies on AI-powered ‘story templates,’ which drill down into databases to produce stories and short reports.
This innovative form of auto-writing has been used for nearly a decade to crank out remarkably precise news stories on sports, finance, real estate and similar data-intensive news.
Plus, the same AI template approach has also been a boon for companies looking to auto-write completely accurate company sales reports, forecasts and the like.
AI as a quick study
In contrast, GPT-3 breaks the AI template tradition by relying on supercomputer-driven predictive analytics. This approach is designed to predict the kind of writing you’re looking for, based on just a few words or few phrases that you input.
GPT-3 is able pull-this-off, given that the AI software has ingested and processed everything published on the Web from 2016 to 2019.
It has also ‘read’ a number of books sourced from numerous libraries.
Essentially, the auto-writing engine comes to users armed with a vast treasure trove of knowledge — as well as an intimate familiarity with a wide spectrum of writing styles.
AI Can Sometimes Get it Wrong
Unfortunately, this genius level of knowledge and writing prowess also comes with a limitation: Given that the entire GPT-3 system is based on sophisticated guessing (AI predictive statistics), the software sometimes ‘guesses’ wrong at the kind of writing you want.
This limitation generally becomes painfully evident when users request a GPT-3 driven app to auto-produce a copy that is 250 words or longer. Too often, those requests are met with longer copy that exhibits questionable logic, nonsequiturs — or even reads as completely nonsensical.
Even so, that limitation has in no way been a deal-breaker. AI programmers have still been able to commercialize dozens upon dozens of GPT-3 driven applications that offer shorter AI-produced copy of sterling quality.
In fact, despite GPT-3’s current limitations, the outlook is extremely bright for increasingly more powerful and more artful auto-writing solutions that sit atop GPT-3.
The Future of AI Powdered Writing
For more than a year, tens of thousands of minds have been tinkering with GPT-3. And many have the goal of coming up with an auto-writing solution that outperforms GPT-3 based auto-writers currently on the market.
Moreover, that number of tinkering minds is expected to increase exponentially, given an OpenAI change in strategy last fall, in which the research company eliminated the GPT-3 waitlist.
Now, any developer in any country supported by GPT-3 can sign-up to develop with GPT-3 — no wait necessary.
Even better: During the past year, a formidable number of competitors have stepped into the arena to challenge GPT-3’s dominance as the major engine of auto-writing — including Cohere, AI21, Microsoft and Nvidia, EleutherAI, and BigScience.
That intense competition virtually guarantees we should see ever more sophisticated auto-writers for businesses and consumers — at extremely attractive price points.